PCI Express is an interconnection technology between chips and an interface technology for extending boards, which employs point-to-point serial interface between devices compared to traditional PCI and earlier shared parallel architectures of the computer bus. Thereby every device is allowed to have its own dedicated and exclusive connection, and needs not to request bandwidth from the whole bus, in addition, by using the connecting characteristics of the serial interface, data transmission speed can be easily increased to a very high frequency which is far more high than the transmission speed of the PCI bus. An individual fundamental PCI express link is a kind of dual simplex link, and a single fundamental PCI express serial interface has two independent links, implemented by different low-voltage to driven signals, consisting of a receive pair and a transmit pair (four lanes). It will not cause conflict with for example USB 2.0, InfiniBand, Ethernet and IEEE 1394/1394b. Therefore, the transmission speed of PCI Express interface is faster than traditional PCI interface, and the PCI Express interface has advantages of such as high-speed serial point-to-point transmission between devices, flexible adjustable bus bandwidth, and the like.
PCI Express consists of four protocol layers: Application Layer, Transaction Layer, Data Link Layer and Physical Layer. And physical layer can be further divided into media access control layer, physical coding sub-layer and physical media access layer.
A PCI Express interface was disclosed in Intel PIPE Macro Rev 1.00 Jun. 19, 2003. The physical coding sub-layer of the PCI express interface is used for implementing data transmission, signal control and error detection to ensure the integrity of transmitted data between physical media access layer and media access control layer.
But the technology of the prior art has defects:
As known from the description of the physical coding sub-layer structure, the physical coding sub-layer can only receive 8-bits input data, but the data inputted to physical coding sub-layer from media access control layer may be 16-bits as well as 8-bits, therefore the data format of physical coding sub-layer in the prior art is not compatible with that of media access control layer.Because sampling errors may occur in the data transmitted from physical media access layer to physical coding sub-layer, but in the description about the structure of physical coding sub-layer in the prior art, there is no component for detection and processing cases of data sampling error, therefore, in prior art, there is high a possibility to transmit wrong data in the physical coding sub-layer.